The University of Luxembourg’s Interdisciplinary Centre for Security, Reliability and Trust (SnT), the Luxembourg Bankers’ Association (ABBL) and Fondation ABBL pour l’éducation financière today announced a new partnership. They will work together to explore and develop trustworthy Artificial Intelligence (AI) tools for instant anti-money laundering due diligence. The project is aimed at crafting a methodology and supporting tools for making AI auditable, traceable and secure.
AI systems are used to learn data models which are then used to make decisions. However, AI tools can acquire biases that can skew their results. For example, AI systems can sometimes ‘learn’ to treat people differently based on their socio-economic status, gender, or place of origin. Biases can be the unintentional result of the way the system is ‘trained’ or coded, or become apparent through malicious intervention. Both scenarios make it crucial to identify and resolve these system errors in order for the AI tools to be reliable when used in financial contexts. Yet checking for these malfunctions is not easy with the naked eye. In many companies, there is no tracking of how AI systems are operated and simple questions like ‘who, how, and when’ regarding a training model are not answered.
These questions become fundamental when related to Anti-Money Laundering (AML) checks, which are key for financial institutions in the context of growing compliance obligations. These vital administrative procedures add complexity and costs. The deployment of instant payment solutions globally has prompted financial institutions to explore trustworthy AI tools suitable for instant AML due diligence.
The importance of ensuring that AI tools are trustworthy has been highlighted through the work that the ABBL and SnT have already been conducting under their established partnership. After consultation with its members, the ABBL has now initiated a new project with SnT, supported by the Fondation ABBL pour l’éducation financière, to develop a solution that banks can apply to their AI based instant AML checks. Throughout the project, SnT researchers will collaborate with ABBL members to ensure that the work is in line with real-world challenges. Interested ABBL members will help with testing the methodology and the proof of concept during this process to ensure the work targets practical business and regulatory challenges.
“AI offers a massive potential to fully automate processes and offer smart assistance to staff. Yet the phenomenon of bias demonstrates the layers of complexity that need to be dealt with before full AI integration can be realised,” said Prof Radu State who leads the project at SnT. “Our long-running partnership with the ABBL and Fondation ABBL pour l’éducation financière is invaluable to us, as it gives our researchers deep insights into the practical challenges faced by organisations working in real world environments.”
“We are embracing technologies such as AI to ensure Luxembourg remains an important global hub for financial services, able to implement new technologies like instant AML with confidence. Our three-year collaboration with SnT is central to these efforts. It connects us to high quality researchers working at the cutting edge of technologies that will be central to our association’s membership for decades to come,” said Ananda Kautz, Head of Innovation, Digital Banking and Payments at the ABBL. “We strive to support the development of high quality, low cost, inclusive banking services, which is why initiatives like this are so important for our members to keep an innovative edge.”